こすたろーんエンジニアの試行錯誤部屋

作成物の備忘録を書いていきますー

【Stable Diffusion】I generated images using the model mutsuki_mix, which can generate realistic human images

スポンサーリンク

I have been trying to generate images using "BlueberryMix" and "BeautyProMix".
This time, I will try to generate images using mutsuki_mix, which is a realistic model and good at generating images of Asians.
Prompts are the same as those used in BlueberryMix and BeautyProMix.
I would be happy to help with model selection and prompting.

contents

abstract

What kind of images can be generated by mutsuki_mix.

1.requirement

Google Colab
Diffusers transformers==4.26.0
model : mutsuki_mix
lora : koreanDollLikeness_v20

2.result

positive

masterpiece, ultra high res, 1 girl, (photo realistic:1.2),small_breasts,cleavage,smile,detailed beautiful skin, face focus, (black eyes), cute, young, looking at viewers, (scoop neck tee:2.0), ,collarbone, ((black hair)), windy, soft light, hand between legs, from front

negative

painting,sketches,(worst quality:2),(low quality:2),(normal quality:2),lowers,normal quality,((monochrome)),((grayscale)),skin spots,acnes,skin blemishes,age spot,(outdoor:1.6),nsfw,ugly face,fat,(missing fingers),extra fingers,extra arms,extra legs,open chest,chest pad

seed = 11527224226351

seed = 6705247472607

seed = 55540575319547

refarence:BlueberryMix

refarence:BeautyProMix

In comparison, BlueberryMix is the closest to Japanese, while BeautyProMix and mutsuki_mix are closer to non-Japanese Asian people.

スポンサーリンク

refarence

civitai.com
github.com
note.com